× To read Now Discovery's COVID-19 Preparedness Plan, please  click here.

The ancient Greek sceptic, Pyrrho of Elis (c. 360 – c. 270 BCE), was the first philosopher to emphasize the frailty of human judgment: We should never believe anything without hard evidence. And since that evidence is hard to come by, he supposed, we should just suspend judgement on everything. Rumour has it that his friends had to follow him around to stop him walking off cliff faces, or into the path of oncoming wagons. This was the only way for him to survive, since he didn’t believe in anything (- including what he saw with his own eyes). 

When working out whether to implement predictive coding during Discovery processes, we need to assess the evidence for this proposal.  

Perhaps the most important case setting out the evidence for predictive coding is the United Kingdom case Pyrrho Investments Limited v MWB Property Limited [2016] EWHC 256 (Ch) (‘Pyrrho Investments’). In this update I set out the evidence in support of predictive coding as made out in that case. 

The 10 Factors in Pyrrho Investments

In Pyrrho Investments, the UK courts first considered the reasonableness of ‘predictive coding’ during ediscovery. In that case, Master Matthews set out 10 factors to explain why predictive coding was reasonable in that situation: 

  1. Other jurisdictions have shown that predictive coding is useful in some cases; 
  2. There is no evidence that predictive coding leads to less accurate disclosure than any other method, and some evidence to the contrary; 
  3. There is greater consistency in using a senior lawyer — plus predictive coding, than using dozens of junior staff to do the job manually; 
  4. There is nothing in the rules that prohibits it; 
  5. The number of electronic documents in that case, to be considered for relevance, was over 3 million; 
  6. The cost of manually searching the documents would be several million pounds at least, and therefore ‘unreasonable’;
  7. Implementing predictive coding solutions would be much cheaper than the manual process; 
  8. The value of the claims being litigated is high – tens of millions. In light of this, the cost of the software is proportionate; 
  9. There is time before the trial for alternative methods to be used if predictive coding doesn’t work out; 
  10. Both parties agree on the use of the software and how it is to be used. 

The Influence of Pyrrho Investments

Master Matthews was clear in Pyrrho Investments that his decision turned on the specifics of the case that came before the bench, and could not be turned into a universal rule. Nevertheless, the decision has been influential in other courts in the United Kingdom and Australia, and is a useful summary of when predictive coding can prove beneficial. 

The wisdom of this approach has been backed up by the explicit support for predictive coding in various forms of court guidance in the UK (Disclosure Review Document, question 14), Australia (SC Practice Note: Technology In Civil Litigation, clause 8.7)  and New Zealand (High Court Rules 2016, Schedule 9). 

Courts appear to be less sceptical about the benefits of predictive coding than many lawyers – and certainly less than some Ancient Greeks. 

About the Author:

Drew Donnelly

Drew Donnelly

Drew Donnelly is a Germany-based legal and tech writer. He was admitted to the Bar in New Zealand, worked as a paralegal in Australia, and comments on eDiscovery from a Europe and Asia-Pacific perspective. Catch him on LinkedIn at https://www.linkedin.com/in/dr-drew-donnelly-b5ba35135/.

Leave a Reply

Your email address will not be published. Required fields are marked *

Interested?
Let’s get started!